Degraded images with local color cast and low contrast can be improved by image quality enhancement techniques considering color, contrast, and other parameters related to a digital image. This article proposes a color image enhancement method which applies weighted multi-scale compensation coefficients to the gray world assumption algorithm based on a color constancy theory. A multi-scale Gaussian filter is used for computing the mean values of the local and global degraded colors, and calculating correction coefficients for size, pixel, and channel of the multi-scale filtered images independently based on the luminance of an image. Then, the weights are determined for a weighted sum of multi-scale correction coefficients by analyzing the local color distribution of the image. Next, the degraded colors are improved by utilizing the correction coefficients, which are integrated into the input image. Finally, the degraded color saturations are improved using the proposed weights. The experimental results show that, compared with conventional methods, the proposed method improves both the color and the contrast of various degraded images and produces better correction results.